blt2114/twisted_diffusion_sampler

Practically and asymptotically accurate conditional sampling from diffusion generative models without conditional training

twisted_diffusion_sampler's Issues

About the motif-scaffolding targets

Hello! Thanks for your awesome work and efforts in open-source the RFDiffusion Motif-Scaffolding Benchmark. I have one question about the provided target structures. I would appreciate it if you can give me some guidance.

The question is that the structures in motif.pdb and target.pdb(e.g. IBCF.pdb, 1BCF_motif.pdb) are not aligned (or not being superimposed). So I can't figure out which part of the target corresponding to the motif. If I understand correctly, the motif should be a part of the target protein and I find in your code you also just took the motif according to the index without any displacements. So I am confused why the motif and the target structures are not superimposed.

Thanks again for your awesome work! Looking forward to your reply~